Volume3,May 2016,

Topic : A Comparative Analysis of Eigen Palm, Fisher Palm and Combined Eigen Palm with Fisher Palm Based Palm Print Authentication

Authors:Vinodkumar || R. Srikantaswamy

Abstract:Personal recognition utilizing palm print, has become a most promising approach recommended by several researchers. Palm print, recognition algorithms are very essentially worthwhile in a wide variety of applications like crime investigation, security control, passport verification etc. This paper describes comparative analysis of palm print, recognition algorithms such as PCA, LDA and combined PCA with LDA. In PCA, the unique images of palm print, are mapped to a minor set of the feature space, which is termed as Eigen palms; they are training set’s eigenvectors and they signify the palm print’s’ principal components pretty best. Formerly, the Eigen palm features will be acquired thru projecting a novel image of palm print, to the subspace which is being spanned by the Eigen palms. In LDA, Every single palm print, image is treated like a coordinate point in higher dimension space of image, which is called palm print, space. Fisher’s linear discriminate is utilized to map palm print, image linearly from this palm print, space into a considerably lesser dimensional space of feature (Fisher palm space), in this space the palm print’s image from the different palm will be discriminated considerably much more proficiently. In combined PCA with LDA: initially we map the palm print, image from image space to Eigen palm space via PCA, furthermore we make use of LDA to attain a classifier which is of linear. The elementary objective of combined LDA and is PCA to advance LDA’s generalization capability. The obtained recognition result from combined PCA with LDA outperforms similar work in the literature including Eigen palms and Fisher palms matching algorithms individually..

Keywords: PCA; LDA; Eigen palms; Fisher palms

Download Paper